[1]

Combustion Synthesized Sodium Manganese (Cobalt) Oxides

as Cathodes for Sodium Ion Batteries

Nicolas Bucher,# Steffen Hartung,

# Irina Gocheva, Yan L. Cheah, Madhavi Srinivasan,

*

Harry E. Hoster*

Abstract

We report on the electrochemical properties of layered manganese oxides, with and without

cobalt substituents, as cathodes in sodium ion batteries. We fabricated sub-µm sized particles

of Na0.7MnO2+z and Na0.7Co0.11Mn0.89O2+z via combustion synthesis. X-ray diffraction

revealed the same layered hexagonal P2-type bronze structure with high crystallinity for both

materials. Potentiostatic and galvanostatic charge/discharge cycles in the range 1.5 V…3.8 V

vs. Na Na+ were performed to identify potential dependent phase transitions, capacity, and

capacity retention. After charging to 3.8 V, both materials had an initial discharge capacity of

138 mA h g-1

at a rate of 0.3 C. For the 20th

cycle those values reduced to 75 mA h g-1

and

92 mA h g-1

for Co-free and Co-doped samples, respectively. Our findings indicate that

earlier works probably underestimated the potential of (doped) P2-type Na0.7MnO2+z as

cathode material for sodium ion batteries in terms of capacity and cycle stability. Apart from

doping a simple optimization parameter seems to be the particle size of the active material.

N. Bucher, S. Hartung, I. Gocheva, M. Srinivasan *, H.E. Hoster *

TUM CREATE

Singapore 138602, Singapore

harry.hoster@tum-create.edu.sg

N. Bucher, S. Hartung, H.E. Hoster *

Technische Universität München

85748 Garching, Germany

Y. L. Cheah, M. Srinivasan *

School of Materials Science and Engineering, Nanyang Technological University

Singapore 639798, Singapore

madhavi@ntu.edu.sg

# Both authors contributed equally to this manuscript

[2]

1. Introduction

Growing interest in the integration of sustainable energy sources into the electricity grid has

spurred immense interest in the field of energy storage applications in recent years. The pre-

dominant commercial battery system is the lithium ion battery (LIB), which has become the

standard energy source for portable electronic devices and currently gains increasing

importance for the electric propulsion of vehicles [1–6]. There are several approaches to

remedy the inherent problems of LIBs, like battery cost and safety. One approach that has

attracted wide attention is aqueous lithium ion battery systems. Only recently, several water-

based batteries with high capacities of approximately 120 mA h g -1

and excellent cycle

efficiencies using LiMn2O4 as cathode material have been developed [7, 8][9]. Another

promising alternative to conventional LIBs is the sodium ion battery (NIB) [10–13] owing to

worldwide abundance of Na metal and lower cost of Na2CO3 (~0.40 € kg-1

) than that of

Li2CO3 (~4 € kg-1

) [14]. Even more important are potential cost savings for the cathode

material. Such cost arguments are even more relevant for larger scale rechargeable batteries

as they are currently discussed for the stabilization of electric grids. Though both Li and Na

are alkali metals, the best electrode materials for LIBs are not necessarily the best for NIBs.

In particular, NIBs offer the potential to avoid the expensive cobalt oxide based cathodes that

are common in LIBs [15]. Two main challenges related to NIBs are (i) the generally lower

voltage and (ii) the larger volume change due to (de-)intercalation of Na+ as compared to Li

+

ions. The latter implies additional strain on the crystal structure during cycling, with negative

impact on the cycle life. As Na based chemistry is less well explored than Li based chemistry

in the field of energy storage, there is a good chance that known structures that do not work

in LIBs are found to work in NIBs.

Layered manganese oxides in their manifold structures [16, 17] are used and discussed as

cathode materials in lithium ion [18, 19] and sodium ion [20, 21] batteries since long. For Li

ion batteries, LiMnO2 with its high theoretical capacity of 285 mA h g-1

is an attractive

alternative to the common LiCoO2, also due to the much lower price of Mn as compared to

Co. An inherent problem of LixMnO2, however, is its instability against a transformation

from a layered into a spinel structure in charged state, i.e., at low Li content x [22], which

goes along with capacity fading. Ma et al. recently highlighted that for NaxMnO2 such a

degradation to a spinel structure is thermodynamically and kinetically disfavoured as

compared to LixMnO2 [23], i.e., NIBs could be fabricated based on this rather cheap and non-

toxic class of cathode materials. Nevertheless, cycling stability remains a central challenge

for Na ion batteries [14, 21] and a broad community is tackling this challenge in several ways

[3]

[24–27]. For layered manganese oxides as cathodes in LIBs, significant improvements of the

cycle life were achieved by partial substitution of Mn by other transition metals

(“doping”)[19]. Interestingly, those cathode materials were prepared via the sodium

containing precursors Na0.7Mn0.89M0.11O2+z (M=Fe, Co, Ni, Cu, Zn, Al; z = 0.05 accounts for

cation vacancies in the MO2 layers [16]), but their electrochemical properties were only

tested after an ion exchange from Na to Li.

In this paper, we will show that the powders generated by the combustion synthesis method

[28] used by Doeff et al. in the cited work [19] are not only useful precursors for the

fabrication of LIB cathodes but are also well-performing cathodes for NIBs. As test cases, we

will show first results for Na0.7MnO2.05 and Na0.7Mn0.89Co0.11O2.05, both with the same P2-

type layered structure [17]. After a description of synthesis and analysis methods and

procedures we will first present the crystal structure and morphology of the samples as

analyzed by X-ray diffraction (XRD) and Field Emission Scanning Electron Microscopy

(FESEM), respectively. We will then evaluate potentiostatic and galvanostatic charge-

discharge curves of both materials with respect to their initial Na content, their potential

depending phases, their cycle efficiency, and their cycling stability. The results will be

discussed in comparison to previous findings for P2-Na0.7MnO2 [20], P2-Na0.6MnO2[21],

monoclinic NaMnO2 [23], P2-NaxCoO2 [29], and P2-Na2/3Co2/3Mn1/3O2 [30].

2. Experimental

2.1 Synthesis

Na0.7MnO2+z was synthesized using the combustion method as described by Doeff et. al. [19,

28, 31, 32]. Sodium nitrate (Sigma Aldrich, ≥ 99%) and Manganese acetate (Alfa Aesar,

anhydrous, 98%) were mixed in Na:Mn mole ratio of 7:10 before being dissolved in

deionized (DI) water, followed by the addition of concentrated nitric acid (≥ 69%,

Honeywell) in excess and 1.5 g of gelatin. The solution was heated to 250 °C until

spontaneous combustion occurred. Oxidation of the amino groups of the gelatin not only

releases the required heat but also yields substantial amounts of gaseous products (N2, CO2,

H2O) that facilitate the reaction and ensure small particle sizes. The product is a dark

brownish powder with cotton-wool-like texture. For better crystallization, the powder was

annealed in air at 800 °C for 4 h. For the cobalt-doped cathode, cobalt nitrate (Sigma Aldrich,

≥ 98%) was added to the precursors in the mole ratio of Na:Co:Mn 7:1.1:8.9. The subsequent

synthesis steps were carried out as described above, obtaining Na0.7Co0.11Mn0.89O2+z

[4]

(according to elemental analysis). In this paper, we will refer to the Co-free and Co-doped

sample as NMO and NCO, respectively.

2.2 Characterization

The crystal structures of the synthesis products were analyzed by powder XRD using a

Bruker X-ray diffractometer with Cu-Kα radiation. 2θ was varied within a range of 10o < 2θ <

80o, using a step size of 0.017

o and 1.7 s per step. The obtained XRD patterns were analyzed

by Rietveld phase analysis using a Lebail approach within the Topas version 3 software,

employing the fundamental parameters approach [33–35]. The powder morphology was

analyzed by FESEM (JEOL JSM-7600F), including elemental analysis. Brunauer-Emmet-

Teller (BET) surface area measurements were carried out with a Nova 3200e surface area and

porosity analyzer.

2.3 Electrochemical measurements

The composite cathodes were prepared by mixing active material (NMO or NCO), acetylene

black (Alfa Aesar, > 99%), and polyvinylidene fluoride (PVdF, Arkema, Kynar HSV 900)

binder in the weight ratio of 60:25:15, with N-Methyl-2-pyrrolidone (NMP, Sigma Aldrich,

≥ 99%), to form a slurry. The well-mixed, homogenous mixture was coated on an Al foil

(16 m in thickness) using a doctor blade, and dried at 80 °C in air to remove the solvent.

The dried coating was then roll-pressed between twin rollers to improve adherence to the Al

foil, and then punched into 16 mm diameter pieces. After drying the electrodes for 4 h under

vacuum at 110 °C, the electrodes were assembled in half-cell configuration in 2016 coin

cells, using 16 mm circular sodium metal pieces as the anode, separated by glass fiber

(Whatman) swollen with sodium-ion conducting electrolyte. That electrolyte consisted of 1M

NaClO4 (Sigma Aldrich, ≥ 98%) in propylene carbonate (PC, Sigma Aldrich, ≥ 99.7%).

Electrochemical studies were carried out via cyclic voltammetry using a BioLogic

potentiostat while galvanostatic testing were performed using a battery tester system

(Neware). All capacities are reported in mA h g-1

, related to the mass of the active material.

For Na0.7MnO2 with a molar mass of 102.41 g mol-1

the extraction/insertion of 1 Na per

formula unit is equivalent to 262 mA h g-1

.

3. Results and Discussion

The crystal structure of P2-Na0.7MnO2 (space group P63mmc, see ref. [17] for the

nomenclature) is depicted in Figure 1. It consists of MnO2 slabs made of edge-sharing MnO6

octahedra. The sodium ions can occupy two different triangular prismatic sites in the plane

[5]

between these MnO2-slabs. Following a nomenclature introduced by Carlier et al. [30] for a

similar system, these are marked as Naf and Nae in Figure 1. For Naf, the triangles of the

prisms are also triangles of the octahedra above and below, whereas the triangles of the

prisms around Nae are formed by edges of three of the octahedral, respectively. Energetically,

the two sites are not equivalent due to their differing number and distance of the Mn3+/4+

ions

surrounding the respective Na+. The lower right corner of Fig.1 exemplarily depicts a

substituting CoO6 octahedron as it appears in NCO. Based on the results of a neutron

diffraction analysis published for P2-Na2/3Co1/3Mn2/3O2 [36], we assume that the lateral

distribution of CoO6 and MnO6 octahedra within the slabs of our Na0.7Co0.11Mn0.89O2+z does

not obey any long-range order. As mentioned above, the MO2 layers have about 1% cation

vacancies [16]. Such a vacancy is illustrated next to the MnO6 octahedron in Fig. 1.

All peaks in the XRD patterns in Figure 2 can be indexed to the layered P2 structure of

Fig. 1. For the lattice vectors a, b, c, a Rietveld analysis yields the lattice parameters a = b

=2.872(2) Å and c = 11.167(2) Å for NMO, and a = b = 2.848(1) Å and c = 11.149(2) Å for

NCO. FESEM images as those in Fig. 3 revealed that both NCO and NMO consist of flat

hexagonal particles after synthesis. This fits to the layered crystal structure detected in XRD.

The lateral dimensions (parallel to lattice vectors a and b) of the flakes in Fig. 3 are in the

range between 300 … 600 nm and their thickness (direction of lattice vector c) lies between

100 nm … 200 nm. Energy-dispersive X-ray spectroscopy (EDX) confirmed Na, Mn, O, and

(for NCO) Co as main constituents of the electrodes. Quantitative analysis revealed a ratio of

cobalt to manganese of 1:6.7, resulting in a stoichiometry of Na0.7Co0.11Mn0.89O2+z for NCO.

Apart from the microstructure the two samples also share the same specific surface area of

~50 m2 g

-1 as measured by BET. In total, the two samples are essentially equal in crystal

structure and morphology and thus provide an ideal model system to separately study the

effect of the Co substituents in NCO on the electrochemical properties.

After assembly the coin cells with either material had an open circuit voltage (OCV) of

2.65 V. This indicates that both cathodes have a comparable Na+ content after synthesis. The

voltage is close to the 2.8 V reported by Caballero for P2-Na0.6MnO2 obtained by a sol-gel

method [21]. We rationalize our slightly lower OCV as a consequence of a higher Na+

content of 0.7 as opposed to 0.6 in ref. [21]. As will be shown more below, both of our tested

materials take up a charge of 56 mA h g-1

when charged up to 3.8 V. This would be

equivalent to an extrusion of 0.21 Na+ per formula unit, i.e., forming Na0.49MnO2+z and

Na0.49Co0.11Mn0.89O2+z, respectively.

[6]

Figure 4 shows cyclic voltammograms (CVs) of NMO and NCO at a scan rate of 0.3

mV s-1

in the range between 1.5 – 3.8 V vs. Na | Na+. We have plotted the second cycle, but

the first and the later cycles are similar. Both CVs are dominated by a couple of redox peaks

between 2 V and 2.5 V. Both peak positions and the apparent equilibrium potential are

slightly more positive for NCO as compared to NMO. A main difference between the two

voltammograms is the presence of manifold oxidation and reduction peaks for NMO in Fig.

4a as opposed to the rather smooth voltammogram of NCO in Fig. 4b. Similar features were

previously shown in plots of the differential specific capacity of P2-Na0.6MnO2 [21], and also

for monoclinic NaMnO2 in potentiostatic intermittent titration (PITT) patterns [23]. These

features can be attributed to phase transitions similar to those intensely studied for P2-

NaxCoO2 [29, 30, 37, 38]. These transitions do not concern the layered P2 backbone structure

itself but the ordering of Na+ and vacancies between the slabs, specifically their distribution

among the two different triangular prismatic sites (see Fig. 1). Apart from the main redox

couple, the most pronounced features in the CV in Fig. 4a are the two oxidation peaks A and

B at 3.5 V and 3.6 V, respectively. These features seem to be a common fingerprint for

potentiostatic and galvanostatic charging of sodium layered oxides: they can be found, e.g., in

the charging curves for P2-Na0.6MnO2 [21], monoclinic NaMnO2 [23], P2-NaxCoO2 [29, 37],

or P2-NaxCo2/3Mn1/3O2 [30], or P2-Nax[Fe1/2Mn1/2]O2 [39]. For P2-NaxCoO2 [29, 37] and P2-

NaxCo2/3Mn1/3O2 [30] this feature was already shown to be related to the transition into/out of

the very stable phase for an Na+ content of 0.5. For P2-Na0.5CoO2, experiments [40, 41] and

calculations [37, 38, 42] revealed an equal occupancy of Nae and Naf sites. Though the

Na+/vacancy patterns for the corresponding Manganese lamellar oxides are not yet known,

similar ordered and stable structures must exist. As a last point related to the CV profiles in

Figs. 4a and 4b we would like to underline the similarity to the charging curves of P2-

NaxCoO2 [29] and P2-NaxCo2/3Mn1/3O2 [30], where partial substitution of Co by Mn “erased”

most of the features apart from the ones around 3.5 V. Our results show that a continuous

increase of the Mn content towards P2-NaxMnO2+z eventually “revives” those features again.

This can be explained as follows: The interaction of Na+ ions with the chemically corrugated

landscape of the doped oxides predominates over the interactions between the Na+ ions. The

latter is mainly responsible for the ordered patterns observed at specific Na+ concentrations.

Even for plain P2-NaxMnO2+z the Na+ ions can be expected to be in a chemically less

homogeneous landscape as compared to P2-NaxCoO2. The cation vacancies in the MnO2

layers (see Fig. 1) [16] act as effective negative charges that trap intercalated Na+ ions [43]

and thus disturb their arrangement into long-range ordered phases. This would explain why

[7]

the voltammetric features observed in Fig. 4a are still relatively weak despite the good

crystallinity of the cathode material.

Fig. 5 presents galvanostatic charge/discharge cycles of NMO and NCO at 0.3 C (50

mA g-1

) between 1.5 – 3.8 V vs. Na | Na+. For NMO, two potential plateaus at 3.5 V and 3.6

V are visible for all cycles. In analogy to the respective features in the CVs in Fig. 4, these

have been marked as A and B. For the second and all further cycles, the charge curves of

NMO exhibit a long plateau at ~2.3 V as dominating features. In agreement to the CVs, the

corresponding discharge plateau is at ~2.1 V. As to be expected from the much broader peaks

in the CVs of NCO, the galvanostatic charge and discharge curves of that sample in Fig. 5b

do not exhibit any pronounced plateaus. NMO and NCO both exhibit a discharge capacity of

138 mA h g-1

after initial charging to 3.8 V. This is equivalent to the intercalation of 0.53 Na+

per formula unit. The capacity is close to the values reported for P2-Na0.6MnO2 (140 mA h

g-1

[21]), P2-NaxCo2/3Mn1/3O2 and P2-NaxCoO2 (both 120 mA h g-1

at C/100 [30]). Only

monoclinic NaMnO2, where up to 0.8 Na+ per formula unit can be reversibly (de-)intercalated

has a significantly higher capacity of 185 mA h g-1

(measured for C/10) [23].

After 20 cycles, NMO and NCO have a remaining capacity of 75 mA h g-1

and 92 mA

h g-1

, respectively. As illustrated in Fig. 6a, the decay in capacity becomes less steep with

increasing cycle number for both cathode materials. This behavior is different from that

reported by Morales et al. for P2-Na0.6MnO2 [21], where the degradation seems to become

faster with increasing number of cycles. Since their crystal structure should be the same as

ours and the voltage range used in their cycles was even narrower, the most likely

explanation for the better stability of our NMO is the electrode structure, specifically, the

particle size. Whereas the flakes generated by our combustion method are all < 1 µm (cf. Fig.

3), the particles in the cited work are > 2 µm. Smaller particles are less sensitive to

mechanical deformation due to Na+ (de-)intercalation. This would explain why we observe a

better stability and it calls for a systematic optimization of the particle size in future studies.

However, as Qu et al. [44] recently reported for α-NaMnO2 used in asymmetric aqueous

supercapacitors, high cycle efficiencies can also be achieved for larger particles. This might

be a hint that part of the fading observed in our work is due to an interaction between the

organic electrolyte and the active material.

An interesting quantity to consider in the context of capacity retention is the cycle

efficiency, i.e., the discharge capacity divided by the preceding charge capacity. Fig. 6b

shows that the cycle efficiencies for both cathode materials are slowly increasing from 78%

in the second cycle to at most 95% in the twentieth cycle. There is no apparent difference

[8]

between NMO and NCO with respect to these values. Similar behaviour, yet not quantified

over many cycles, was reported for P2-Na0.6MnO2 [21], monoclinic NaMnO2 [23], P2-

NaxCo2/3Mn1/3O2 [30], and P2-NaxCoO2 [29]. The main contribution to the low coulomb

efficiency comes from side reactions at higher potentials, e.g., electrolyte decomposition [13,

23]. Our results do not show any effect of Co-doping on these parasitic reactions. Another

possible reason for capacity fading could be the disproportionation of Mn3+

into Mn4+

and

dissolved Mn2+

, which is a general problem of Mn containing cathodes. This hypothesis is

substantiated by literature [1, 6] and is a key issue to be addressed in future works.

The better capacity retention for the doped sample fits to corresponding observations

for the stability of substituted layered manganese oxides in lithium ion batteries [32]. It has to

be pointed out, however, that the ion exchange of Na+ towards Li

+ goes along with a

transition of a P2 to an O2 structure [32], so that the structural boundary conditions cannot be

directly compared. Moreover, the capacity fading of LIBs involving LixMnO2 cathodes is

mainly due to a gradual transition into an energetically favourable spinel crystal structure [45,

46]. This transition does not occur in NIBs [23]. The stabilizing effect of Co could involve

more than one factor, similar to what was observed for spinel LiCoyMn2-yO4 as compared to

LiMn2O4 as cathodes in LIBs where Co was shown to improve the general chemical stability

and the cation conductivity of the electrode [6].

4. Conclusions

The gelatin assisted combustion method is ideal for the synthesis of P2-type layered

sodium manganese oxides as well-defined model cathodes for rechargeable Na+ ion batteries.

These cathodes can be systematically varied by partial substitution of Mn by Co, Fe, Ni, Cu,

Zn, Al. The resulting powders consist of sub-µm flakes with a P63mmc crystal structure.

Being charged to 3.8 V vs. Na | Na+, the cathodes based on those particles have a discharge

capacity of 140 mA h g-1

, equivalent to the extrusion and re-insertion of ~0.5 Na. Due to the

small size of the flakes, these capacities were attainable at a rate of 0.3 C. We have

exemplarily shown that substitution by Co does not alter the capacity itself but has a

beneficial effect on the cycling stability. Specifically, the capacity of the 20th

discharge cycle

(with 0.3 C) increases from 75 mA h g-1

for Na0.7MnO2+z to 92 mA h g-1

for

Na0.7Co0.11Mn0.89O2+z. An effect of Co doping towards improved cycle efficiency via

suppression of parasitic side reactions, however, can be excluded. Cyclic voltammograms

indicate that the presence of Co suppresses the formation of ordered Na+ / vacancy phases

which is likely to have an influence on the Na+ transport within the cathode. Further studies

[9]

will reveal in how far this behaviour directly contributes to the observed cycle life

improvement. Moreover, a systematic study on the optimum particle size for plain and doped

NaxMnO2 is on its way.

Acknowledgements

The authors would like to thank Miss Han-Yi Chen, Miss Yin Ting Teng and Mr. Jan Geder

for their valuable help in conducting SEM, XRD, and BET measurements. This work was

financially supported by the Singapore National Research Foundation under its Campus for

Research Excellence and Technological Enterprise (CREATE) programme.

[10]

References

1. Wu Y, Wan C, Jiang C, Fang S (2002) Introduction, Principles and Advances of

Lithium Secondary Batteries. Tsinghua University Press, Beijing

2. Scrosati B, Garche J (2010) J Power Sources 195:2419–2430

3. Liu C, Li F, Ma L, Cheng H (2010) Adv Mater 22:E28–62

4. Ceder G, Hautier G, Jain A, Ong SP (2012) MRS Bulletin 37:185–191

5. Vetter J, Novák P, Wagner MR, Veit C, Möller K-C, Besenhard JO, Winter M,

Wohlfahrt-Mehrens M, Vogler C, Hammouche A (2005) J Power Sources 147:269–

281

6. Wu YP, Rahm E, Holze R (2002) Electrochim Acta 47:3491–3507

7. Liu L, Tian F, Zhou M, Guo H, Wang X (2012) Electrochim Acta 70:360–364

8. Qu Q, Fu L, Zhan X, Samuelis D, Maier J, Li L, Tian S, Li Z, Wu Y (2011) Energy

Environ Sci 4:3985

9. Tang W, Liu LL, Tian S, Li L, Li LL, Yue YB, Bai Y, Wu YP, Zhu K, Holze R (2011)

Electrochemistry Communications 13:1159–1162

10. Ellis BL, Nazar LF (2012) Curr Opin Solid State Mater Sci 16:168–177

11. Chevrier VL, Ceder G (2011) J Electrochem Soc 158:A1011–A1014

12. Ellis BL, Makahnouk WRM, Makimura Y, Toghill K, Nazar LF (2007) Nat Mater

6:749–753

13. Kim S-W, Seo D-H, Ma X, Ceder G, Kang K (2012) Adv Energy Mater 2:710–721

14. Palomares V, Serras P, Villaluenga I, Hueso KB, Carretero-González J, Rojo T (2012)

Energy Environ Sci 5:5884–5901

15. Tarascon J-M, Armand M (2001) Nature 414:359–367

16. Parant J-P, Olazcuaga R, Devalette M, Fouassier C, Hagenmuller P (1971) J Solid

State Chem 3:1–11

17. Delmas C, Fouassier C, Hagenmuller P (1980) Physica B+C 99:81–85

18. Doeff MM, Anapolsky A, Edman L, Richardson TJ, Jonghe LC De (2001) J

Electrochem Soc 148:A230–A236

19. Eriksson TA, Lee YJ, Hollingsworth J, Reimer JA, Cairns EJ, Zhang X, Doeff MM

(2003) Chem Mater 15:4456–4463

[11]

20. Mendiboure A, Delmas C, Hagenmuller P (1985) J Solid State Chem 57:323–331

21. Caballero A, Hernán L, Morales J, Sánchez L, Santos Peña J, Aranda MAG (2002) J

Mater Chem 12:1142–1147

22. Shao-Horn Y (1999) J Electrochem Soc 146:2404–2412

23. Ma X, Chen H, Ceder G (2011) J Electrochem Soc 158:A1307–A1312

24. Kim H, Kim D, Seo D, Yeom M, Kang K, Kim DK, Jung Y (2012) Chem Mater

24:1205–1211

25. Tarascon J, Guyomard D, Wilkens B (1992) Solid State Ionics 57:113–120

26. Sauvage F, Laffont L, Tarascon J-M, Baudrin E (2007) Inorg Chem 46:3289–3294

27. Yang S, Wang X, Wang Y, Chen Q, Li J, Yang X (2010) T Nonferr Metal Soc

20:1892–1898

28. Chick LA, Pederson LR, Maupin GD, Bates JL, Thomas LE, Exarhos GJ (1990) Mater

Lett 10:6–12

29. Berthelot R, Carlier D, Delmas C (2011) Nat Mater 10:74–80

30. Carlier D, Cheng JH, Berthelot R, Guignard M, Yoncheva M, Stoyanova R, Hwang

BJ, Delmas C (2011) Dalton Trans 40:9306–12

31. Dollé M, Hollingsworth J, Richardson TJ, Doeff MM (2004) Solid State Ionics

175:225–228

32. Dollé M, Patoux S, Doeff MM (2005) Chem Mater 17:1036–1043

33. Rietveld HM (1969) J Appl Crystallogr 2:65–71

34. Rietveld HM (1967) Acta Crystallogr 22:151–152

35. Cheary RW, Coelho A (1992) J Appl Crystallogr 25:109–121

36. Lu Z, Donaberger R a., Dahn JR (2000) Chem Mater 12:3583–3590

37. Shu G, Prodi A, Chu S, Lee Y, Sheu H, Chou F (2007) Phys Rev B: Condens Matter

76:184115

38. Meng YS, Hinuma Y, Ceder G (2008) J Chem Phys 128:104708

39. Yabuuchi N, Kajiyama M, Iwatate J, Nishikawa H, Hitomi S, Okuyama R, Usui R,

Yamada Y, Komaba S (2012) Nat Mater 11:512–517

40. Huang Q, Foo ML, Lynn JW, Zandbergen HW, Lawes G, Wang Y, Toby BH, Ramirez

AP, Ong NP, Cava RJ (2004) J Phys Condens Matter 16:5803–5814

[12]

41. Zandbergen H, Foo M, Xu Q, Kumar V, Cava R (2004) Phys Rev B: Condens Matter

70:024101

42. Zhang P, Capaz R, Cohen M, Louie S (2005) Phys Rev B: Condens Matter 71:153102

43. Robertson AD, Armstrong AR, Bruce PG (2001) Chem Mater 13:2380–2386

44. Qu QT, Shi Y, Tian S, Chen YH, Wu YP, Holze R (2009) J Power Sources 194:1222–

1225

45. Wohlfahrt-Mehrens M, Butz A, Oesten R, Arnold G, Hemmer RP, Huggins RA (1997)

J Power Sources 68:582–585

46. Robertson AD, Armstrong a. R, Paterson AJ, Duncan MJ, Bruce PG (2003) J Mater

Chem 13:2367–2373

[13]

List of Figures

Fig. 1 Crystal structure of P2 layered NaxMnO2 and NaxCoyMn1-yO2. Two possible sites for

intercalated Na and three possible ways of filling the octahedral sites (vacancy, Mn, Co) are

shown (see text).

Fig. 2 XRD patterns of (a) NMO and (b) NCO. (c) Rietveld refinement of XRD pattern of

NCO and (d) hkl positions of space group P63mmc.

Fig. 3 FESEM images of NMO (a) and NCO (b); inset in (a) illustrates the layered structure.

Fig. 4 Cyclic voltammograms of a NMO and b NCO (scan rate: 0.3 mV s-1

).

Fig. 5 Galvanostatic charge / discharge profiles of a NMO and b NCO at 0.3 C. 1st, 10

th, and

20th

cycle are shown. Inset: derivative of the charging curves in the range 3.45 V – 3.7 V for

better visibility of features A and B.

Fig. 6 a Integrated charge of galvanostatic discharge at 0.3 C for 20 cycles for both cathode

materials; b Coulomb efficiencies (discharge capacity divided by preceding charge capacity)

for cycles 2…20.

[14]

Fig. 1 Crystal structure of P2 layered NaxMnO2 and NaxCoyMn1-yO2. Two possible sites for

intercalated Na and three possible ways of filling the octahedral sites (vacancy, Mn, Co) are

shown (see text).

Bucher et al., Figure 1

[15]

Fig. 2 XRD patterns of (a) NMO and (b) NCO. (c) Rietveld refinement of XRD pattern of

NCO and (d) hkl positions of space group P63mmc.

Bucher et al., Figure 2

10 20 30 40 50 60 70 80

2 Theta

a

b

c

d

0 0

2

0 0

4

0 1

0

0 1

2

0 1

3

0 0

6

0 1

1

0 1

6

0 1

1

0 0

8

0 1

4

1 1

2

1 1

4

0 2

00

2 1

0 1

8

P2-Na0.7MnO2+z (NMO)

P2-Na0.7Co0.11MnO2+z (NCO)

2

[16]

Fig. 3 FESEM images of NMO (a) and NCO (b); inset in (a) illustrates the layered structure.

Bucher et al., Figure 3

[17]

Fig. 4 Cyclic voltammograms of a NMO and b NCO (scan rate: 0.3 mV s-1

).

Bucher et al., Figure 4

1.5 2 2.5 3 3.5

(b)

0.4

0.3

0.2

0.1

0

-0.1

-0.2

-0.3

(Potential vs. Na|Na+) / V

Cu

rren

t /

mA

P2-Na0.7Co0.11MnO2+z

(NCO)

(a)

0.4

0.2

0.1

0

-0.1

-0.2

-0.3

1.5 2 2.5 3 3.5

0.3

P2-Na0.7MnO2+z

(NMO)

A B

A B

Cycle 1Cycle 2Cycles 3 - 5

0.5

0.5

-0.4

-0.4

Cycle 2

Cycle 3 - 5

Cycle 1

[18]

Fig. 5 Galvanostatic charge / discharge profiles of a NMO and b NCO at 0.3 C. 1st, 10

th, and

20th

cycle are shown. Inset: derivative of the charging curves in the range 3.45 V – 3.7 V for

better visibility of features A and B.

Bucher et al., Figure 5

0 20 40 60 80 100 120 140 160

0 20 40 60 80 100 120 140 160

Capacity / (mA h g-1)

(Po

ten

tial

vs.

Na|

Na+ )

/ V

4.0

3.5

3.0

2.5

2.0

1.5

4.0

3.5

3.0

2.5

2.0

1.5

3.8 V

1.5 V

3.8 V

1.5 V

1st 10st20st

1st10st20st

1st10st20st

1st

(a)

(b)

P2-Na0.7MnO2+z

(NMO)

P2-Na0.7Co0.11MnO2+z

(NCO)

10st20st

AB

1st

3.5 3.6Potential / V

A B

dPo

ten

tial

/ d

Cap

acit

y(a

.u.)

1st

10st

20st

[19]

Fig. 6 a Integrated charge of galvanostatic discharge at 0.3 C for 20 cycles for both cathode

materials; b Coulomb efficiencies (discharge capacity divided by preceding charge capacity)

for cycles 2…20.

Bucher et al., Figure 6

Cycle number

Co

ulo

mb

Eff

icie

ncy

Dis

char

ge C

apac

ity

/ (m

A h

g-1

)

P2-Na0.7MnO2+z

(NMO)

P2-Na0.7Co0.11MnO2+z

(NCO)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20180

160

140

120

100

80

60

40

20

0

100 %

90 %

80 %

70 %

60 %

50 %